Neal Mohan, CEO of YouTube, said that using videos on the platform to train artificial intelligence (AI) models would "clearly violate" YouTube's terms and conditions, after OpenAI's CTO said he "didn't know" if the tool used YouTube videos for training.

In an interview with Bloomberg, Mohan made his first public statement about the Sora tool, which caused a sensation and excitement when it was announced earlier by OpenAI.

Mohan said, "From the perspective of creators, when they upload their hard work to our platform, they have certain expectations. One of those expectations is that the terms of service will be followed. It does not allow downloading of content such as transcripts or video clips, which clearly violates our terms of service. These are the rules regarding content on our platform."

In March of this year, Mira Murati, CTO of OpenAI, discussed this new generative AI tool in an interview with The Wall Street Journal. The tool can create videos lasting several minutes. When asked about the training data for Sora, Murati said, "We used publicly available data and licensed data," but the CTO did not know if this included content from YouTube, Instagram, and Facebook.

OpenAI's training data has raised questions

OpenAI has been secretive about the training data used in creating large language models and other generative AI tools, which has led to multiple lawsuits.

Comedian Sarah Silverman and a group of other authors have filed a lawsuit against OpenAI under California's unfair competition law, accusing the company of using copyrighted materials in its training data.

The New York Times has also filed a copyright infringement lawsuit against the AI company, seeking accountability for the damages caused by OpenAI's unauthorized use of copyrighted materials.

According to a recent report by The Wall Street Journal, OpenAI plans to use YouTube video transcripts to train ChatGPT-5, indicating that they are not hindered by these lawsuits.